Generating alternative solutions for dynamic programing models of water resources problems

For a decision maker to bring implicit knowledge, that is, knowledge that cannot be incorporated into a model, to bear on a problem, he must be confronted with alternatives that are perceivable as different with respect to this implicit knowledge. A technique is presented to generate alternatives that are different from each other, but good with respect to modeled objectives, for problems that can be modeled by dynamic programing. The technique is argued to be better than alternative techniques. An application to floodplain management shows that the technique can produce sets of different alternatives for water resources problems.